Friday, February 13, 2015

High Tech "State of the Union", Q4 2014 - Part 3


Once more upon the waters! yet once more!
And the waves bound beneath me as a steed
That knows his rider. - Lord Byron

The story of The Seven Voyages of Sinbad was first introduced to western readers in the 19th century by Sir Richard Burton, the famed explorer, geographer and orientalist, from his translation into English of "The Book of 1,001 Nights." The tome is sometimes referred to as 1,001 Arabian Nights, as Burton's translation was of an Arabic text. However, the stories narrated by Scheherazade to her husband the king during those long nights over nearly three years drew from a wide range of tales originating from or inspired by legends and folkore from Arab, Persian, Indian and even Ancient Greek sources.

I read these tales as a youngster when my insatiable appetite for books first emerged - a hunger which has never abated. The section pertaining to Sinbad's seven voyages was particularly enchanting. I could imagine myself feeling the waves surging under the keel of my dhow, triangle sail full and billowing, as the vessel cut its way across an emerald sea with a cloudless blue sky above. That feeling is very beautifully captured in the following musical score by Nikolai Rimsky-Korsakov (and though the recording is nearly 50 years old, it is still my favorite rendition of the piece:)

This week, our own adventures on the stormy, windswept oceans of High Tech continue, as we examine the financial results of Q4 2014 for our remaining sector groupings - the Carolingians, who represent the flower of the European semiconductor industry, and the Stone Masons, whose claim to fame is worldwide leadership in SoC technology.

The Carolingians


Yet the best pilots have need of mariners, besides sails, anchor and other tackle. - Ben Jonson

Europe's destiny has always been intimately linked to the sea. Many of its momentous events and historical turning points were centered on nautical episodes: Salamis, Actium, Lepanto, the Spanish Armada in the Channel, Trafalgar and Jutland, to name a few.

The Age of Discovery was a maritime manifestation of the Renaissance as it broke out of Italy and spread across the European continent. Following on the heels of Columbus and Vespucci - both Italian navigators sponsored by Spanish and Portuguese monarchs, respectfully - the nations of western Europe scrambled to establish trade routes and outposts to the Far East and the New World. 

Populous nations such as France, England and Spain found themselves competing against each other for the power and riches that these explorations promised, while tiny nations such as Portugal and Holland leaped at the opportunity to break out of their geographic limitations to establish world-spanning territorial and commercial empires. The smallest countries were the most aggressive of all, as the Dutch colonized islands from the Caribbean to Indonesia, with additional holdings in North and South America, southern Africa & India while the Portuguese sent missionaries to Japan, traders and colonists to Macau & Goa and grabbed a huge chunk of South America, calling it Brazil and establishing the first empire upon which the sun never set. Naval warfare became an endemic problem of global trade as seafaring powers old & new fought each other directly with great Men O' War and Ships of the Line bristling with cannon and indirectly thru commissioned privateers as they contested each other's control of the sea lanes.

The strength and vitality of a civilization is measured by its impetus for exploration and discovery. These adventures into the unknown are the living embodiment of a society's dreams and aspirations and reflect on its innate curiosity and potential for cultural, social and economic growth.

So how are the Carolingians doing as they explore the frontiers of High Tech in their quest to find what lies over the Technology horizon for Europe? Let's see what story the numbers recount to us.

At first glance, the companies are all strikingly similar in organization and stated goals. As a group, they are vertically integrated firms that by and large eschew the consumer electronics and wireline networking markets in favor of opportunities in automotive and ISM (industrial, scientific and medical) sectors. The attraction to automotive is obvious, as the already substantial electronic content in new car models is expected to grow considerably as automotive-related IoT applications mature and disseminate. ISM is, on the other hand, highly fragmented and typically low margin business. Nevertheless, the hope of these and other companies focused on the segment is that they might position themselves as dominant players in enough sub-sectors of ISM so that they could rely on that portion of the business as a 'cash cow' to provide some level of income stability for the company.

As a consequence of their general market orientation, the Carolingians have focused on enriching their engineering expertise in fields that are less mainstream. To be specific, they are somewhat less interested in digital applications and more focused on adding value to their offerings thru developing greater acumen in analog, mixed signal, power, security and wireless design. This, as it fortuitously turns out, is a potentially explosive combination, as it is these very capabilities which will be of greatest importance to the fledgling IoT market.

This, however, is where the similarities between the companies end. NXP is clearly the strongest of the three, at least in terms of momentum. It has achieved three years of steady growth and is arguably better positioned for the IoT than its rivals. The company is making a sustained and substantive effort to be a major player on world markets as well as a leading MCU provider - a crucial advantage in IoT positioning. 

Infineon is close behind and can rely on the special loyalty of its German customer base. However, for two of the last three years the company's revenues have been stagnant. The firm is at this point trailing NXP in total revenue, earnings growth & momentum and is now forced to chase after its rival. 

STMicro is the ailing company of the trio. Once a very adventurous firm that showed tremendous mastery in developing unique value-add and new technology, the continuous decay in revenues over the last four years is incontestable proof that this ship has a broken rudder and a complacent (if not an outright oblivious) leadership. 

None of the many partnerships STMicro has engaged in over the years has borne fruit. This suggests that the company either doesn't know how to gain advantage from partnerships or it doesn't really care. Stated differently: There appear to be no real mariners aboard the once proud ship, as STMicro's allegedly strategic actions seem to have no substance and the company appears to have no long term strategic direction or the tactics to support it. 

In these last three weeks of High Tech company analyses, only IBM seems to be as adrift as STMicro. The corrective action, therefore, should be identical - the shareholders need to incite a mutiny and the officers of the company must be made to walk the plank.

The Stone Masons

And as great seamen, using all their wealth
And skills in Neptune's deep invisible paths,
In tall ships richly built and ribbed with brass,
To put a girdle round about the world. - George Chapman

This grouping is the cream of the crop in the semiconductor industry - true Ships of the Line. Qualcomm has been the dominant force in mobile telephony for over two decades. Broadcom is, arguably, the inventor of the SoC and holds strong market positions in broadband, enterprise networking and wireless. Mediatek has a very strong presence in China and is a leader in mobile baseband. NVidia has near monopolistic dominance in graphics and is pursuing a wide range of applied R&D initiatives with its parallel processing GPU technology.

Despite the differences between these firms in terms of individual strengths and market focus, they share certain fundamental characteristics. They have global reach and extraordinarily rich IP portfolios. Their understanding of system level hardware and software is tremendous, allowing them to extract value from their system OEM customers. One would look to these companies to be on the bleeding edge of......well, everything.

How are they each performing, then, both individually and collectively, as they are buffeted by the gusts and waves on a stormy High Tech sea? Let's see what the numbers say (Note: Mediatek's numbers, normally reported in New Taiwan dollars, are converted in the graph to USD using a ratio of 32 NTD/$.)

As a group, the performance of the Stone Masons is decidedly mixed. If the overall global economy were truly healthy, all the revenue numbers would be trending up and to the right. Yet the results are, overall, rather mediocre.

Qualcomm is by far the strongest of the four, having crushed all its competitors in the smartphone market, including Broadcom and Nvidia in the application processor segment. The flattening of the growth curve for smartphones and Apple's market share capture last quarter should have led to a continuation and perhaps worsening of Qualcomm's two year growth taper - but Q4's number shows that it didn't. What's going on?

The winds and waves are always on the side of the ablest navigators. - Edward Gibbon, "The Decline and Fall of the Roman Empire"

The current revenue cycle brings a great deal of bad news for Qualcomm. Firstly, margins are dropping. This was to be expected as price becomes more of a competitive factor in the saturating & plateauing smartphone market. To complicate matters, Qualcomm is having a great deal of difficulty collecting license payments from customers in China. It appears that some OEMs - especially Chinese 'white box' manufacturers of mobile phones - are deliberately under-reporting volume shipments while others might even be refusing payment outright. 

Source: wikipedia

The Chinese government has multiple departments and governmental agencies - most associated with the military - deeply embedded in the nation's High Tech industry. Furthermore, despite a rigorously comprehensive legal codification of intellectual property protection, China's court system is so poor at defending IP rights that even domestic companies consider it pointless to use it as a forum to defend their intellectual property. 

Consequently, Qualcomm has found itself in the same predicament as many other western companies that have preceded it in the China market, participating in a marketplace that behaves as a government-sanctioned cartel of coordinated theft & abuse of IP combined with murderous price pressure. It is, in effect, a new and unique model of economic warfare, with the theatre of conflict being the massive potential of China's 1.4B consumers serving as bait for a trap to lure Western industrial and technology firms with the promise of fabulous riches while extracting fiscal and IP endowments as the price of entry to play in the Middle Kingdom's huge sandbox. Stated differently: western companies are effectively subsidizing their own future competitors in world markets with the false promise of fair access to the Chinese domestic market. The bedazzling temptation of China's domestic market potential has been so great that it has prevented Western industrial segments and their governments from organizing and implementing an effective coordinated defense.

It would be of course silly to grandstand about the "moral turpitude" of such an industrial policy. After all, economic warfare is as old as mankind and there is not a tribe, people or country which has not engaged in it. Beijing is simply playing the game and cleverly so as they pursue their national interests. As a practical matter, one must simply recognize the situation faced by Qualcomm (and others) in order to come to terms with it and deal with the circumstances effectively. 

Clearly the Chinese government has a firm grasp of its leverage in this state of affairs - the China National Development and Reform Commission (NDRC) has doubled down on Qualcomm and poured salt in its wounds thru filing anti-trust charges against the company. Qualcomm paid a $975M fine to settle the antitrust case and in return now has legal sanction to force collection of the delinquent invoices. Yet one cannot help but wonder if the fine is near the same value as the total of payments owed to it by Chinese customers.The NDRC further forced Qualcomm to lower its royalty rates by 35%. Part of the deal also compels Qualcomm to build more of its products with SMIC, a chip fab sponsored by Beijing which has long been a particularly weak player in the merchant foundry business. 

The unhappy news for Qualcomm doesn't end there. Samsung is refusing to employ the SnapDragon 810 in its next line of smartphones, citing heat problems that other OEMs have also protested. Perhaps Qualcomm can do something to ameliorate this in deeper process nodes, but the sub-28nm quandary facing the chip industry today - wherein the 3P's cannot all be simultaneously improved - poses difficult technical problems as well as painful financial choices for the company's future.

Finally: following up from last quarter's report, Qualcomm has also failed to announce any plans to deal with the glove thrown quite publicly by Intel, who declared that there would be no quarter given and no effort spared to penetrate the mobile computing market. In keeping with the theme of this editorial series, we can think of Intel as having raised the Jolly Roger on the mainmast, proclaiming itself a buccaneer on the high seas of system-on-chip and openly targeting the smartphone and tablet sea lanes. With semiconductor process, CPU and 3D-IC technology second to none, Intel's Man O'War sports a cannonade of potentially legendary proportions. Qualcomm has yet to publicly respond to the challenge, and the longer it takes, the more worrisome will be its outlook.


Yet not all the news is bad. Interestingly, Qualcomm's baseband revenue is growing faster than the applications processor line. This might be due to continuing troubles at Mediatek, whose normally market-leading baseband business has been losing share of late to Marvell. Perhaps Mediatek's weakness is greater than the market generally realizes. Also, Apple's iPhones use Qualcomm baseband chips - an oddity, since everyone else prefers integrated applications and baseband processors for system power and cost benefits. Nonetheless, this serves to mitigate the damage to Qualcomm's app processor revenues from Apple's conquest of a significant chunk of Android market share with the iPhone 6. Finally, Qualcomm is, as recommended in the last quarterly analysis, moving forward with an ARM architectural license to design the company's own 64b CPU core, expected to prototype before the end of the year in an as yet unannounced node.

My own view is that Qualcomm has both the superlative technical muscle and the executive management acuity to overcome its (admittedly unnerving) set of difficulties and beat its opponents in the gathering zero sum game struggle for market share in a stagnating mobile computing market. For those of you who have been following this blog for a while, I fully realize that my approval of a given executive team is so uncharacteristic that you are undoubtedly wondering if I've spent too much time up in the crow's nest of the mainmast without a hat. Yet one must acknowledge that the numbers speak for themselves - Qualcomm grew substantially last quarter in the face of a hurricane that would have broken the masts and torn the sails from almost any other company.

Yet my respect and enthusiasm for the company is muted by two factors. First of all, Qualcomm needs to play offense with China. The company is large and strong enough that it should covertly begin in earnest the work of building a coalition of non-Chinese governments and companies to take organized action against what is, at the moment, a unidirectional war being waged by Beijing for worldwide High Tech leadership. It is, simply put, a game that two can play and is a matter of cold calculation and pragmatism for non-Chinese governmental and corporate entities to fully acknowledge and deal with the threat with both a solid defense and, more importantly, an effective offense.

Furthermore, Qualcomm needs to more openly display its technical firepower, both against Intel's challenge and in terms of its preparedness to compete effectively in the IoT. The executive mindset of Qualcomm appears to me the most sanguine & down-to-earth and least cluttered by trite bromides with regards to  the IoT's current status and potential. Reiterating last quarter's view, what I await in particular from Qualcomm is the announcement of an MCU version of Snapdragon targeted at IoT applications.
Moving on.....

Nvidia is another bright spot in the Stone Mason portfolio, albeit much dimmer than mighty Qualcomm. The company's sales grew in 2014 from $4.1B to nearly $4.7B - a 14% increase and more than twice the semiconductor market growth rate (excluding memories.) 

Yet despite this impressive revenue performance, the company remains a one-trick pony. Put another way - it looks like a three-masted brig, but carries only one line of cannon instead of the two or three that would be expected of a true ship of the line. Fully 80% of Nvidia's sales are GPUs for the PC market. The Tegra smartphone applications processor line is making inroads into automotive, but overall sales continue to decline steeply, contributing only 10% to the company's revenues. Excluding the quarterly $66M royalty from Intel for licensed graphics technology, the remaining contributions from Nvidia's other business lines is more or less background noise.

There are some very, very capable technical people at Nvidia. I know a few of them and would work with them again at the drop of a hat. Yet after years of R&D in a multitude of areas - high performance computing, ISM, automotive and others, all of it based on its extraordinarily powerful array of parallel graphics processors - the company appears unable to substantively capitalize on its investments in technology innovation.

In this sense, Nvidia appears to resemble Google in many ways - lots of great research into new applications but little to show for it in return. The reason for this, though, is very different from Google's. Nvidia is widely known for its fierce embrace of the worst of Silicon Valley management practices - "drive 'em 'til they drop, then sell the corpses to the soap factory", "the whippings will continue until morale improves" and so forth. The company demands results but apparently leaves no room for people to stick their necks out and take risks, fail and thereupon learn from their failures.

A leader is best when people barely know he exists; Not so good when people obey and acclaim him; worst when they despise him; but a good leader who talks little when his work is done, his aim fulfilled, they will say "We did it our selves!" - Laozi, "Tao Te Ching"

Drive matters in reaching success for any endeavor. High Tech has always attracted people with nerve and drive in abundance. Yet most High Tech firms value hierarchy and control over initiative and spirit. Stated differently: senior and executive management finds it much easier to manage then to lead. 

There's a very large executive staff at Nvidia. Yet the strength of an organization does not spring from how many VP's a company has and their ability to crack the whip to instill dread of oversight and fear of punishment, but by its ability to bring forth the talents, drive and willing, eager commitment of its people toward the goals and aspirations of the enterprise. 

If you see one organization who's general manager regularly barks at his employees in meetings, bellowing demands for greater results and more work output, his workers blanching at the sight of him while carefully picking every vague phrase and noncommittal remark they utter, then behold another organization where its workers are clearly chomping at the bit and their GM has his hands full just holding onto the reins, barely restraining them from running riot, which do you think is more likely to succeed? Mike Ditka, head coach of the Chicago Bears from 1982-1992, and Chuck Noll, head coach of the Pittsburgh Steelers from 1969-1992, would pick the latter group. Who do you think Blackbeard would pick? How about Black Bart? Barbarossa? Calico Jack?


The damaging effects of an autocratic executive regime are even more apparent with Broadcom. Nvidia is trying to develop new business lines and can't; Broadcom, on the other hand, has multiple business lines but can't do anything with them. 

The numbers tell the tale. Broadcom was a $7.5B company in 2011; in 2014, it was $8.4B. That's 4% growth/year - barely ahead of the most optimistic inflation estimates. The company has strength in wireless, networking and broadband (particularly STB.) Its IP portfolio is arguably the deepest and broadest of the group.Yet no matter how Broadcom's officers set their sails, it is the current carrying them along - not the wind.

Broadcom prides itself on hiring 'only the best.' Without question, the company is as rich in talent as it is in IP. Yet this is meaningless if the executive staff doesn't know how to bring the best out of that talent pool.

The company is beginning to approach increasingly rougher seas with an organizational structure and philosophy that is not prepared to deal with foul weather. As described in an earlier post for this series, Intel's pirate flag has been spotted on the mobile computing sea lanes, but it has also apparently flown covertly thru the normally placid waters of Broadcom's STB business - and word is that Intel sees an opportunity to plunder a rich haul of pieces of eight. 


This brings us to our final member of the Stone Masons - Mediatek. The company has long demonstrated strength in consumer electronics and mobile telephony, particularly with its baseband products. It remains the strongest player in the China market for baseband and strengthened its DTV business thru the acquisition of MStar. Revenue growth in 2014 was primarily from revenue recognition of the merger, with resulting financials demonstrating the wisdom of the acquisition.

Yet revenue growth tapered and then, surprisingly, rolled over in the second half of the year. Marvell has been digging into Mediatek's baseband revenue in China. Furthermore, despite Mediatek's claim that it took 15% of Qualcomm's 4G LTE baseband business in the Middle Kingdom during 2014, Qualcomm's worldwide baseband business grew - and not at the expense of Marvell.

To reassure stockholders, Mediatek announced that it intends to grow its baseband line outside of China, which currently represents about 2/3 of the business, so that the China portion drops to only half of its total worldwide revenues. The question remains whether this proportional change in market presence will be of the company's own volition or will reflect a loss of China's market share to its rivals.

"Damn the torpedoes, full speed ahead!" - Admiral David Farragut, Union Navy, at the Battle of Mobile Bay, Alabama, American Civil War, 1864

The company clearly has an aggressive long term strategic vision, though. It intends to recruit another 2,000 technical professionals to expand its R&D efforts. Furthermore, Mediatek's success against Qualcomm in 4G LTE suggests that they do not plan to cede the high end to anyone. The struggle for dominance in China's smartphone market promises to be an energetic one, with banners flying and cannons roaring.



Next week, in the final installment of the series, we will have to bring our ship into a suitable port in the Gulf of Corinth, as the final phase of our journey takes us overland to the slopes of Mount Parnassus so that we may bring offerings to the Pythia.....